-
Notifications
You must be signed in to change notification settings - Fork 634
Closed
Closed
Copy link
Labels
bugSomething isn't workingSomething isn't working
Description
Snakemake version
v8.18.2
Describe the bug
When running snakemake locally, only one job is submitted even though there are enough resources for more.
Logs
Initially there are 500Gb, 80 cores and 42 jobs ready:
Resources before job selection: {'mem_mb': 500000, '_cores': 80, '_nodes': 9223372036854775807, '_job_count': 9223372036854775807} Ready jobs (42)
and each jobs requires 1Gb and 1 core:
resources: tmpdir=/scratch/production/temp/, mem=1 GiB, mem_mb=1024, runtime=60, shell_exec=sh
However, only one job is selected to run:
Selected jobs (1)
even though there are still available resources:
Resources after job selection: {'mem_mb': 498976, '_cores': 79, '_nodes': 9223372036854775806, '_job_count': 1}
Execute 1 jobs...
Additional context
Looking into the results from the ILP solver, only one has math.isclose(variable.value(), 1.0) == True.
It happens the same with the greedy scheduler, only one job is selected.
EDIT: the problem is with the new option --max-jobs-per-second introduced on v8.17.0. Since it has no default, it seems it just submits one job at a time.
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
bugSomething isn't workingSomething isn't working